A demo dataset of 18 rigged humanoid character assets generated for multi-modal training. Each asset was produced from a single 3D model using Gemini for image generation, Tripo3D for conversion and auto-rigging, and the Metrixel CLI to output multiple data types. The dataset was created by EntVista and last updated on Hugging Face in May 2026.
Use Cases
- Train multi-view 3D reconstruction models based on the static multi-view turntable renders.
- Develop signed-distance-field (SDF) volume estimation techniques based on the provided SDF volumes.
- Benchmark auto-rigging and 3D asset generation pipelines based on the described workflow.
Strengths
- Contains 18 distinct rigged humanoid character assets.
- Each asset provides multiple data modalities, including renders and SDF volumes, as described.
Limitations
- Description metadata is limited; actual data quality requires manual inspection after download.
- Row count is unknown, which may limit suitability assessment.
- Column-level documentation is absent; field semantics must be inferred after download.
Provenance
- Source
- EntVista via Hugging Face.
- Collection Method
- Generated from a single 3D model using Gemini, Tripo3D, and the Metrixel CLI (version 2026.2.0+4384).
- Freshness
- Last updated 2026-05-23 05:35:42; freshness should be verified.